The Edinburgh Post-natal Depression Scale (EPDS) was validated on a community sample of 702 women at six weeks post-partum using Research Diagnostic Criteria for depression. The estimates of sensitivity, specificity and positive predictive value, being based on a large random sample, offer improved guidelines for the use of the EPDS by the primary care team.

No abstract

The autosomal dominant syndrome of Hereditary Nonpolyposis Colorectal Cancer (HNPCC) is due to germline DNA mismatch repair gene mutations in most cases. However, the penetrance of such mutations outwith classical HNPCC kindreds is unknown because families studied to date have been specifically selected for research purposes. Using a population-based strategy, we have calculated the lifetime cancer risk associated with germline DNA mismatch repair gene mutations, irrespective of their family history. We identified 67 gene carriers whose risk to age 70 for all cancers was 91% for males and 69% for females. The risk of developing colorectal cancer was significantly greater for males than for females (74% versus 30%, P= 0.006). The risk of uterine cancer (42%) exceeded that for colorectal cancer in females, emphasising the need for uterine screening. Our findings give further insight into the biological effect of defective DNA mismatch repair. We have demonstrated a systematic approach to identifying individuals at high risk of cancer but who may not be part of classical HNPCC families. The risk estimates derived from these analyses provide a rational basis on which to guide genetic counselling and to tailor clinical surveillance.

The difficulty of identifying susceptibility genes for common diseases has polarized geneticists' views on what disease models are appropriate and how best to proceed once high-density genome maps become available. Different disease models have different implications for using linkage or linkage-disequilibrium-based approaches for mapping complex disease genes. We argue that the choice of study population is a critical factor when designing a study, and that genetically simplified isolates are more useful than diverse continental populations under most assumptions.